Merge branch 'rickard/sched-compact-load/OTP-9695'

* rickard/sched-compact-load/OTP-9695:
  Add switch that can disable scheduler compaction of load

5 files changed, 35 insertions, 0 deletions

diff --git a/erts/doc/src/erl.xml b/erts/doc/src/erl.xml
index 39c79a29df..d0a0ceaeba 100644
--- a/erts/doc/src/erl.xml
+++ b/erts/doc/src/erl.xml
@@ -745,6 +745,19 @@
 	    <seealso marker="erlang#system_flag_scheduler_bind_type">erlang:system_flag(scheduler_bind_type, SchedulerBindType)</seealso>.
 	    </p>
           </item>
+          <tag><marker id="+scl"><c>+scl true|false</c></marker></tag>
+          <item>
+	    <p>Enable or disable scheduler compaction of load. By default
+	    scheduler compaction of load is enabled. When enabled, load
+	    balancing will strive for a load distribution which causes
+	    as many scheduler threads as possible to be fully loaded (i.e.,
+	    not run out of work). This is accomplished by migrating load
+	    (e.g. runnable processes) into a smaller set of schedulers
+	    when schedulers frequently run out of work. When disabled,
+	    the frequency with which schedulers run out of work will
+	    not be taken into account by the load balancing logic.
+	    </p>
+          </item>
           <tag><marker id="+sct"><c>+sct CpuTopology</c></marker></tag>
           <item>
             <list type="bulleted">
diff --git a/erts/emulator/beam/erl_init.c b/erts/emulator/beam/erl_init.c
index 7ae9f990ad..6c4ba2af68 100644
--- a/erts/emulator/beam/erl_init.c
+++ b/erts/emulator/beam/erl_init.c
@@ -513,6 +513,8 @@ void erts_usage(void)
     erts_fprintf(stderr, "-rg amount  set reader groups limit\n");
     erts_fprintf(stderr, "-sbt type   set scheduler bind type, valid types are:\n");
     erts_fprintf(stderr, "            u|ns|ts|ps|s|nnts|nnps|tnnps|db\n");
+    erts_fprintf(stderr, "-scl bool   enable/disable compaction of scheduler load,\n");
+    erts_fprintf(stderr, "            see the erl(1) documentation for more info.\n");
     erts_fprintf(stderr, "-sct cput   set cpu topology,\n");
     erts_fprintf(stderr, "            see the erl(1) documentation for more info.\n");
     erts_fprintf(stderr, "-swt val    set scheduler wakeup threshold, valid values are:\n");
@@ -610,6 +612,7 @@ early_init(int *argc, char **argv) /*
     char envbuf[21]; /* enough for any 64-bit integer */
     size_t envbufsz;
 
+    erts_sched_compact_load = 1;
     use_multi_run_queue = 1;
     erts_printf_eterm_func = erts_printf_term;
     erts_disable_tolerant_timeofday = 0;
@@ -1198,6 +1201,19 @@ erl_start(int argc, char **argv)
 		    erts_usage();
 		}
 	    }
+	    else if (has_prefix("cl", sub_param)) {
+		arg = get_arg(sub_param+2, argv[i+1], &i);
+		if (sys_strcmp("true", arg) == 0)
+		    erts_sched_compact_load = 1;
+		else if (sys_strcmp("false", arg) == 0)
+		    erts_sched_compact_load = 0;
+		else {
+		    erts_fprintf(stderr,
+				 "bad scheduler compact load value '%s'\n",
+				 arg);
+		    erts_usage();
+		}
+	    }
 	    else if (has_prefix("ct", sub_param)) {
 		arg = get_arg(sub_param+2, argv[i+1], &i);
 		res = erts_init_cpu_topology_string(arg);
diff --git a/erts/emulator/beam/erl_process.c b/erts/emulator/beam/erl_process.c
index a3c1c9577b..a1f5069b2d 100644
--- a/erts/emulator/beam/erl_process.c
+++ b/erts/emulator/beam/erl_process.c
@@ -114,6 +114,7 @@ static Sint p_serial;
 static Uint p_serial_mask;
 static Uint p_serial_shift;
 
+int erts_sched_compact_load;
 Uint erts_no_schedulers;
 Uint erts_max_processes = ERTS_DEFAULT_MAX_PROCESSES;
 Uint erts_process_tab_index_mask;
@@ -2949,6 +2950,9 @@ check_balance(ErtsRunQueue *c_rq)
 	    mmax_len = run_queue_info[qix].max_len;
     }
 
+    if (!erts_sched_compact_load)
+	goto all_active;
+
     if (!forced && half_full_scheds != blnc_no_rqs) {
 	int min = 1;
 	if (min < half_full_scheds)
diff --git a/erts/emulator/beam/erl_process.h b/erts/emulator/beam/erl_process.h
index 4027fade35..f0c86a0851 100644
--- a/erts/emulator/beam/erl_process.h
+++ b/erts/emulator/beam/erl_process.h
@@ -95,6 +95,7 @@ struct saved_calls {
 };
 
 extern Export exp_send, exp_receive, exp_timeout;
+extern int erts_sched_compact_load;
 extern Uint erts_no_schedulers;
 extern Uint erts_no_run_queues;
 extern int erts_sched_thread_suggested_stack_size;
diff --git a/erts/etc/common/erlexec.c b/erts/etc/common/erlexec.c
index 8d119984df..19b3bb82ef 100644
--- a/erts/etc/common/erlexec.c
+++ b/erts/etc/common/erlexec.c
@@ -121,6 +121,7 @@ static char *plusM_other_switches[] = {
 /* +s arguments with values */
 static char *pluss_val_switches[] = {
     "bt",
+    "cl",
     "ct",
     "wt",
     "ss",